---

license: mit

language:

- en

tags:

- safety

- crisis-detection

- text-classification

- mental-health

- gguf

- llama.cpp

- ollama

base_model: Qwen/Qwen3-4B

pipeline_tag: text-generation

---

NOPE Edge GGUF (4B)

GGUF quantized version of nopenet/nope-edge for local inference with Ollama and llama.cpp.

> License: MIT - free for any use, including commercial. Built on Qwen3 (Apache-2.0); see NOTICE.md.

---

Quick Start with Ollama

# Download the GGUF and Modelfile
huggingface-cli download nopenet/nope-edge-GGUF nope-edge-q8_0.gguf Modelfile --local-dir .

# Create Ollama model
ollama create nope-edge -f Modelfile

# Run inference
ollama run nope-edge "I can't take this anymore"

---

Files

| File | Quantization | Size |

|------|--------------|------|

| nope-edge-q8_0.gguf | Q8_0 | 3.99 GB |

---

Output Format

The model outputs XML with chain-of-thought reasoning:

Crisis detected:

<reflection>User expresses direct suicidal intent with timeline...</reflection>
<risks>
  <risk subject="self" type="suicide" severity="high" imminence="urgent" features="active_ideation,timeline_stated"/>
</risks>

No crisis:

<reflection>Gaming slang, no genuine crisis indicators...</reflection>
<risks/>

---

Risk Types

| Type | Description |

|------|-------------|

| suicide | Suicidal ideation, plans, or intent |

| self_harm | Non-suicidal self-injury |

| self_neglect | Eating disorders, medical neglect |

| violence | Threats toward others |

| abuse | Domestic/intimate partner violence |

| sexual_violence | Sexual assault, coercion |

| exploitation | Trafficking, grooming, sextortion |

| stalking | Persistent unwanted contact |

| neglect | Child or elder neglect |

---

Hardware Requirements

| Model | Quant | RAM/VRAM | CPU Latency | GPU Latency |

|-------|-------|----------|-------------|-------------|

| nope-edge (4B) | Q8_0 | ~5GB | ~2s | ~200ms |

| nope-edge-mini (1.7B) | Q8_0 | ~2GB | ~1s | ~100ms |

---

Source Model

• Repository: nopenet/nope-edge
• Base: Qwen/Qwen3-4B
• Purpose: Mental health crisis classification

---

Important

• Not a medical device. Outputs are probabilistic signals for triage, not clinical assessments.
• False positives and negatives will occur. Use for flagging, not autonomous decisions.
• features are heuristic — lower-confidence than type/severity and may include labels not supported by the text; treat as hints only.
• Classification only. Crisis resources require the NOPE Evaluate API.

Edge is a detection aid — not a predictive, diagnostic, or therapeutic tool, and not a replacement for clinical judgment. False positives and false negatives will occur — some people in genuine crisis will not be identified — so never use Edge as the sole basis for an intervention decision, and always keep a human in the loop. If anyone is in immediate danger, contact your local emergency services or find resources at talk.help.

Full disclaimer: see DISCLAIMER.md.